Variability and Changes in the North Atlantic Oscillation Index

Abstract

Most – or perhaps even all – of the observed variations in the winter North Atlantic Oscillation (NAO) index can be explained as internally-generated climate variability. The influence of external forcing factors on the observed NAO behaviour is still an open question. Two sea level pressure datasets yield different results for the strength of the winter NAO trend from the 1960s to the 1990s, though these trends from both datasets lie outside the 90% range of trends generated by the internal variability of climate models, and the latter is more precisely known now that over 8000 years of simulation under constant forcing is available for analysis. Similarly, a much expanded, multi-model ensemble of simulations under increasing anthropogenic forcing strengthens earlier findings of a shift in the mean state of the winter atmospheric circulation towards positive NAO conditions. There is considerable inter-model spread in both the magnitude of this response to increased forcing and in its regional structure, but of the 21 climate models analysed here, none showed an overall decrease in the mean level of the NAO index. If a shift towards positive NAO conditions is a realistic response to increasing anthropogenic forcing, then this signal has not yet emerged from the natural variability: observations since the 1990s show a return to lower values, and the 2009/2010 winter had the record negative NAO index in a record lasting almost two centuries. It is possible that anthropogenic forcing could be altering the temporal or spatial character of the interannual NAO variability, though only relatively small changes in pattern are evident when considering the multi-model ensemble as a whole and there is only weak evidence for an increase in the interannual variance.